Sony is making a PlayStation Network login optional for several of its single-player games on PC. The company says that you'll no longer need one for God of War Ragnark, Horizon Zero Dawn Remastered, Marvels Spider-Man 2 (which arrives on Steam and the Epic Games Store this Thursday) or The Last of Us Part II Remastered (which hits PC on April 3).However, those who do sign in with a PSN account while playing those titles will get some in-game goodies such as cosmetic outfits, early unlocks of certain gear and other resources. They'll also be able to unlock trophies and have access to a friends management feature.For months, players have been protesting the need to link their PSN account to their Steam or Epic Games Store purchases to play single-player games. It at least to some degree makes sense for multiplayer elements of games like Ghost of Tsushima. But as evidenced by Sony swiftly dropping the requirement from Helldivers 2 after a major backlash from players, the PSN account linking strategy has by and large been a mess.Meanwhile, Sony has revealed more details about what you can expect from Marvels Spider-Man 2 on PC. The port from Nixxes includes multiple ray-tracing options, along with upscaling and frame generation. The game also has pretty modest minimum specs, so it might still run decently on a laptop that you've been rocking for several years.This article originally appeared on Engadget at https://www.engadget.com/gaming/pc/sony-is-ditching-the-psn-login-requirement-on-some-of-its-pc-games-183259677.html?src=rss

No Man's Sky is back with Worlds Part II, expanding the universe in new ways as only games can. Hello Games, the studio behind No Man's Sky, has announced the release of its newest iteration and the billions of stars and trillions of planets added to its universe.In a statement about the release, Sean Murray, founder of Hello Games, explained, "This allows us to push the boundaries of our engine and technology without changing the things people love about the game already. If you settled on your home planet with a beautiful base that you lovingly crafted, that is safe but now there are new worlds to explore with a level of variety no one has seen before."This shift includes the addition of Gas Giants, which Hello Games claims are up to 10 times bigger than any planets previously included in the game. The developers have also improved things under the sea, with better fishing and deep sea diving. Plus, the lighting should be better, with more detailed shadows, sharper sunlight and, of course, those shining starry skies.Hello Games also says Worlds Part II offers quests that connect existing storylines and mysteries. A lot of this development comes from work on Hello Game's new fantasy game Light No Fire.This article originally appeared on Engadget at https://www.engadget.com/gaming/playstation/no-mans-sky-update-adds-billions-of-stars-and-even-more-planets-165102362.html?src=rss

Social engineering on a trusted platform works well for threat actors, McAfee says.

The first Dungeons & Dragons VR trailer reveals gameplay, but theres no release date yet.

Entropy Studio, one of the largest VFX companies to have existed for over two decades, pitched in for the sequel Culpa Tuya to the global hit Culpa Ma. Their mission was to generate the high-octane race sequence, and they delivered a highly effective visual scene that set the pace for an explosive third part.The team is meticulous in replacing standard cars filmed on the set with photorealistic luxury vehicles to maintain seamless consistency in lighting, reflections, and interactions using matchmoving and integration. The team of around 50 specialists was necessary to complete over 180 shots for a 10-minute sequence, which included technical issues, such as speed variations and camera adjustments.Adding to the show, Entropy Studio was responsible for a 3D drone design and simulating a destructible digital bus stop shelter as well as crafting collision effects. Coached by VFX Supervisor Ramn Cervera, along with 3D Supervisor Pedro Cabrera, the result is a breath-taking sequence that has cemented Culpa Tuya to be a visual masterpiece.The post Behind the VFX of Culpa Tuya: Crafting the Thrilling Race Sequence appeared first on Vfxexpress.

Two years ago, Volkswagen debuted a prototype of its ID.7, the sixth model in its lineup of electric vehicles, at the Consumer Electronics Show in Las Vegas. The automaker said it was planning to launch the EV in three primary markets: Europe, China, and North America. But now, after a series of delays, VW has canceled its plans to sell the ID.7 in North America.Due to the ongoing challenging EV climate, Volkswagen has decided to no longer offer the ID.7 to the North American market, a Volkswagen spokesperson said. Despite this, electric vehicles continue to be a core part of Volkswagens long-term product strategy, and new electric models will continue to be introduced for this market.EV sales in the U.S. have been growing. They jumped more than 15% in the last quarter of 2024, compared to the year prior, setting a new quarter sales record. Sales for the entire year reached 1.3 million in the U.S., a more than 7% increase compared to 2023, according to data from Cox Automotive.But the Trump administration has already made the EV climate more challenging. Moments after being sworn in this week, Sean Duffy, Donald Trumps new Transportation Secretary, signed a memorandum to roll back fuel economy standards with the aim of eliminat[ing] the electric vehicle mandate.This follows an executive order by Trump promising to do the same, though no such mandate to force Americans to purchase EVs actually exists. What Trump is actually targeting are the EV tax credits of up to $7,500 that are part of the Inflation Reduction Act and Bidens goal to have EVs make up 50% of new vehicle sales by 2030.[Photo: Volkswagen]Some automakers were already scaling back their EV plans. Ford announced in August 2024 that it was canceling some electric SUVs in favor of hybrid models as a way to improve profits, though said it was still planning to roll out new all-electric vehicles in the future. Volkswagens ID.7 was first planned for 2024, but last year the automaker announced an indefinite delay due to changing market dynamics. The ID.7 has already launched in both Europe and China.Despite solid sales for 2024, automakers are now on edge as the EV industry in the U.S. faces an uncertain future. Trumps anti-EV agenda and plans for tariffs pose a threat to EV sales. The legality of Trumps actions are yet to be seen, however. And even still, analysts do expect EV sales to grow in 2025. Cox Automotives 2025 outlook predicted that one out of every four vehicles sold in 2025 will be electrified. The market is gaining momentum, economic fundamentals are improving, and consumer sentiment is pointing in the right direction, Cox Automotive chief economist Jonathan Smoke said in a statement. We are ready for what 2025 might bring.

Spirit Airlines has rejected a merger offer with Frontier as it prepares to exit bankruptcy.Wednesday, Frontier made its second offer to merge with the bankrupt Spirit Airlines, but Spirit rejected it on the grounds that it was financially insufficient.In 2022, Frontier offered to acquire Spirit for $2.9 billion, but the offer was ultimately rejected when Spirit chose to accept a higher offer from JetBlue (which was later blocked for antitrust concerns).Frontier Airlines put forward its current merger offer in hopes of creating a strong, low-fare airline together.We have long believed a combination with Spirit would allow us to unlock additional value-creation opportunities, said Barry Biffle, CEO of Frontier, in a statement.In a joint letter to Spirits chair and CEO, Biffle and Frontiers chair of the board added that they believe the transaction generates more value for all Spirit stakeholders than Spirits current plan filed to the Bankruptcy Court.But Frontiers offer was lower than the amount the two parties had discussed in 2022, Raniero DAversa, an attorney and market-leading practitioner in bankruptcies, out-of-court restructurings, and creditors rights controversies, tells Fast Company.In Frontiers offer, debt holders would receive $400 million in new debt and 19% of Frontiers common equity. It would also require stakeholders to invest $350 million in equity, which they were not willing to do, according to a regulatory filing.The offer appears to be too little, too late, DAversa says.In its rejection of the offer, Spirit said that the board believes Frontiers proposal is so insufficient as not to merit a counter.Accepting or considering this offer could also interfere with the airlines plans to exit bankruptcy, which it had filed for in November. Spirit Airlines is on a fast track to exit, DAversa says.Any serious consideration of the Frontier offer would derail the whole bankruptcy process, which is overwhelmingly supported by its constituents. The airline has a February 13 court date to finalize its exit plan.While a company and its board have a fiduciary obligation to consider any deal for the benefit of its constituents and equity holdersand in the case of bankruptcy, its creditorstheyre under no obligation to actually accept it.Although Spirit has operated normally during its bankruptcy, the airline has cut 200 jobs and sold some Airbus planes in order to raise millions of dollars.DAversa compares the situation to the classic a bird in the hand . . . idiom. Spirit is lined up to come out of bankruptcy imminently, and it appears the airline is ready to fight through its final month rather than merge.

Despite initiating a probe into Chinese AI startup DeepSeek, Microsoft has added the startups latest reasoning model R1, to its model catalog on Azure AI Foundry and GitHub.As part of theblog postin which Microsoft declared that it has added R1, it said that the model had undergone rigorous red teaming and safety evaluations, including automated assessments of model behavior and extensive security reviews to mitigate potential risks.

Thursday, January 30th, 2025Posted by Jim ThackerCheck out free Android and iOS 3D scanning app RealityScan 1.6html PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/REC-html40/loose.dtd"#RealityScan 1.6 is hereTransform everyday objects into high-quality 3D assetsright from your phone!New features to explore:Advanced Camera Control ModeNew Processing Options PanelDownload RealityScan 1.6 for free on your favorite app store!#RealityScan #3DScanning pic.twitter.com/2voda575o2 Capturing Reality (@RealityCapture_) January 28, 2025Epic Games has released RealityScan 1.6, the latest version of its free photogrammetry app for iPhones, iPads and Android devices.The update provides more control over the scanning process, with users able to generate 3D models with up to 1 million polygons and 8K textures.Free app makes photogrammetry available to anyone with a smartphoneFirst released in 2022, RealityScan is intended to make the 3D scanning capabilities of tools like Epic Games RealityCapture accessible to hobbyists as well as pros.The app turns photos of objects captured using the camera in a mobile phone or tablet into textured 3D models for use in AR, game development or general 3D work.Models are generated online in standard 3D file formats, and can be used in DCC apps and game engines like 3ds Max, Blender, Cinema 4D, Maya, Unity and Unreal Engine.RealityScan 1.6: more manual control over the scanning processRealityScan 1.6 introduces new options designed to provide more user control over scans.Launching a project in the new Camera Control mode gives full manual control over camera settings, making it possible to adjust focus, shutter speed, ISO values, white balance and flash.The original Augmented Reality mode provides a more guided experience, with a real-time display of the point cloud being generated, overlaid with camera positions.In addition, the new Processing Options panel makes it possible to choose processing options, including the resolution and filtering method for the 3D mesh and textures generated.Users can choose to generate meshes with up to 100,000 polygons or up to 1 million polygons, and with 4K or 8K textures.System requirements and availabilityRealityScan is compatible with Android phones and tablets that support ARCore, running Android 7.0+, and with iPhones and iPads running iOS 16.0+ and iPadOS 16.0+. The software is free, and can be used on personal projects, including commercial work. By default, the EULA gives Epic Games the right to use your scan data to train products and services, but you can opt out in the in-app settings.Read a full list of new features in RealityScan 1.6Download free Android 3D scanning app RealityScan from Google PlayDownload free iOS 3D scanning app RealityScan from the App StoreHave your say on this story by following CG Channel on Facebook, Instagram and X (formerly Twitter). As well as being able to comment on stories, followers of our social media accounts can see videos we dont post on the site itself, including making-ofs for the latest VFX movies, animations, games cinematics and motion graphics projects.Latest NewsCheck out free Android and iOS 3D scanning app RealityScan 1.6Update to Epic Games' free app provides more control over scans, with the option to create 1-million-poly models with 8K textures.Thursday, January 30th, 2025Chaos releases V-Ray 7 for HoudiniCheck out the new features in the renderer, including a new Houdini-specific volume shader and better integration with Solaris.Wednesday, January 29th, 2025Chaos releases V-Ray 7 for MayaCheck out the new features in the VFX renderer, including OpenPBR support and changes to USD support specific to the Maya edition.Wednesday, January 29th, 2025Adobe releases Substance 3D Modeler 1.19Check out the latest features in the VR modeling app, including the new Assets workflow, Shell parameter, and updates to Mesh to Clay.Wednesday, January 29th, 2025Red9 ProPack gets new tools created for Sonic the Hedgehog 3Maya animation toolkit for VFX and games gets new tools developed for Red9's character rigging work on the latest Sonic movie.Tuesday, January 28th, 2025JangaFX teases IlluGenThe EmberGen dev's new app will be a 'better way to create the majority of assets' for VFX in games. Here's what we know about it so far.Monday, January 27th, 2025More NewsCheck out the new features due in EmberGen 2.0Tutorial: Creating Runtime Cinematics in Unreal Engine 5PhotoLine 25 now reads Cryptomatte dataPolygonflow releases Dash 1.8.5After Effects' latest beta lets you preview HDR compsNew version of iOS and Android scanning app Kiri EngineRE:Vision Effects launches Twixtor StandaloneSee the new features due in Unreal Engine 5.6 and beyondLeft Angle releases Autograph 2025Check out this GPU-accelerated Houdini Skin Slide DeformerCheck out this free 2D Rim Light Tool for NukeCreate motion graphics in Nuke with the new Screen FX pluginOlder Posts

Post-training quantization (PTQ) focuses on reducing the size and improving the speed of large language models (LLMs) to make them more practical for real-world use. Such models require large data volumes, but strongly skewed and highly heterogeneous data distribution during quantization presents considerable difficulties. This would inevitably expand the quantization range, making it, in most values, a less accurate expression and reducing general performance in model precision. While PTQ methods aim to address these issues, challenges remain in effectively distributing data across the entire quantization space, limiting the potential for optimization and hindering broader deployment in resource-constrained environments.Current Post-training quantization (PTQ) methods of large language models (LLMs) focus on weight-only and weight-activation quantization. Weight-only methods, such as GPTQ, AWQ, and OWQ, attempt to reduce memory usage by minimizing quantization errors or addressing activation outliers but fail to optimize precision for all values fully. Techniques like QuIP and QuIP# use random matrices and vector quantization but remain limited in handling extreme data distributions. Weight-activation quantization aims to speed up inference by quantizing both weights and activations. Yet, methods like SmoothQuant, ZeroQuant, and QuaRot struggle to manage the dominance of activation outliers, causing errors in most values. Overall, these methods rely on heuristic approaches and fail to optimize data distribution across the entire quantization space, which limits performance and efficiency.To address the limitations of heuristic post-training quantization (PTQ) methods and the lack of a metric for assessing quantization efficiency, researchers from the Houmo AI, Nanjing University, and Southeast University proposed the Quantization Space Utilization Rate (QSUR) concept. QSUR measures how effectively weight and activation distributions utilize the quantization space, offering a quantitative basis to evaluate and improve PTQ methods. The metric leverages statistical properties like eigenvalue decomposition and confidence ellipsoids to calculate the hypervolume of weight and activation distributions. QSUR analysis shows how linear and rotational transformations affect quantization efficiency, with specific techniques reducing inter-channel disparities and minimizing outliers to enhance performance.Researchers proposed the OSTQuant framework, which combines orthogonal and scaling transformations to optimize large language models weight and activation distributions. This approach integrates learnable equivalent transformation pairs of diagonal scaling and orthogonal matrices, guaranteeing computational efficiency while preserving equivalence at quantization. It reduces overfitting without compromising the output of the original network at the time of inference. OSTQuant uses inter-block learning to propagate transformations globally across LLM blocks, employing techniques like Weight Outlier Minimization Initialization (WOMI) for effective initialization. The method achieves higher QSUR, reduces runtime overhead, and enhances quantization performance in LLMs.For evaluation purposes, researchers applied OSTQuant to the LLaMA family (LLaMA-1, LLaMA-2, and LLaMA-3) and assessed performance using perplexity on WikiText2 and nine zero-shot tasks. Compared to methods like SmoothQuant, GPTQ, Quarot, and SpinQuant, OSTQuant consistently outperformed them, achieving at least 99.5% floating-point accuracy under the 4-16-16 setup and significantly narrowing performance gaps. LLaMA-3-8B incurred only a 0.29-point drop in zero-shot tasks, compared to losses exceeding 1.55 points for others. In harder scenarios, OSTQuant was better than SpinQuant and gained as much as 6.53 points by LLaMA-2 7B in the 4-4-16 setup. The KL-Top loss function provided a better fitting of semantics and reduced noise, thus enhancing performance and lowering gaps in the W4A4KV4 by 32%. These outcomes showed that OSTQuant is more effective at outlier handling and ensuring distributions are more unbiased.In the end, the proposed method optimized the data distributions in the quantization space based on the QSUR metric and the loss function, KL-Top, improving the performance of large language models. With low calibration data, it diminished noise and preserved semantic richness compared to existing quantization techniques, achieving high performance in multiple benchmarks. This framework can serve as a basis for future work, starting a process that will be instrumental in perfecting quantization techniques and making models more efficient for applications requiring high computation efficiency in resource-constrained settings.Check out the Paper. All credit for this research goes to the researchers of this project. Also,dont forget to follow us onTwitter and join ourTelegram Channel andLinkedIn Group. Dont Forget to join our70k+ ML SubReddit. [Recommended Read] Nebius AI Studio expands with vision models, new language models, embeddings and LoRA (Promoted)The post Quantization Space Utilization Rate (QSUR): A Novel Post-Training Quantization Method Designed to Enhance the Efficiency of Large Language Models (LLMs) appeared first on MarkTechPost.